OpenAI's new tiered model lineup (Sol at $5/$30, Terra at $2.50/$15, Luna at $1/$6 per million tokens) is live today, and based on GPT-5.5's 47-point citation swing at launch, GEO operators should expect measurable source shifts within 48 hours.
Why it matters: New model releases are now effectively citation volatility events — marketing operators should baseline their brand's AI visibility before the rollout settles or the shift will read as a mystery drop.
Google confirmed its highest-ever query volume after Argentina's July 7 match, proving real-time and breaking-news queries remain a category where AI Mode can't safely summarize fast enough.
Why it matters: Event-driven and time-sensitive queries are still AI Overview dead zones — brands in those sectors retain durable organic territory worth investing in.
Instead of routing users directly to a site, Google is inserting a google.com/goto redirect on search result clicks, which alters referrer strings in analytics and changes how organic click data is captured.
Why it matters: If this rolls out broadly, GA4 referral source attribution will shift — flag it before teams misread a data change as a real traffic drop.
Two independent analyses caught ChatGPT swapping its primary search backend (Labrador, Bright, Oxylabs, SERP) between identical runs, with 11.6% of prompts citing different sources on a re-query — making prompt-level GEO measurement unreliable.
Why it matters: Winning consistent citations in ChatGPT requires presence across multiple retrieval layers — optimizing for a single surface is not a durable strategy.
E-commerce operators can now use validFrom/validThrough and Product.category to align structured data with Merchant Center feed specs, making pricing windows and product taxonomy machine-readable for AI Shopping.
Why it matters: Low-effort structured data update that positions e-commerce pages for the AI Shopping layer Google is actively expanding — a quick technical win worth adding to any site audit.